1 December 2018 in Physical Communication
17 November 2022 in IEEE Internet of Things Journal
Non-Orthogonal Multiple Access (NOMA) is classified into grant-based (GB) RA and grant-free (GF) RA in 5G systems. GF NOMA RA allows UEs to directly transmit data packet over uplink resources without grant information so that the signaling overhead is substantially reduced. This approach is suitable for small-size data packet IoT services. However, power collision interference resulting from uncoordinated resource selection leads to serious transmission failure in massive MTC scenarios. As a result, the number of transmissions and access delay are sharply increased. In this article, we make the best of the advantages of both GB RA and GF RA and propose a hybrid grant (HG) NOMA RA scheme to address the signaling overhead and power collision issues. The objective of the proposed scheme is to achieve an acceptable tradeoff among signaling overhead, access failure probability and access delay. The UE selects the access resource without uplink grant, and the BS assists to avoid inter-NOMA power collision by limited interference resolution (IR) signaling. To comprehensively analyze the performance, the closed-form expression of outage probability is derived by adopting the order statistic theory and the sum of independent exponential distributions, and the maximum number of transmissions and access delay are theoretical analyzed. A Monte Carlo simulation framework is built to verify the theoretical analysis. Finally, the analytical and simulation results show that HG NOMA RA outperforms GB NOMA RA and GF NOMA RA in massive access scenarios.
4 December 2023 in IEEE Sensors Journal
Behavior recognition technologies based on sensor devices show strong development potential as sensor devices are increasingly applied in people’s lives. However, some challenges in sensor-based behavior recognition, such as incomplete feature extraction, complex behavior recognition algorithms and the recognition accuracy, remain open issues. To address these issues, this paper proposes an improved deep convolutional long short-term memory (DeepConvLSTM) model that uses the squeeze and excitation (SE) module and group convolutions to perform human activity recognition (HAR). The proposed scheme uses convolutional neural network (CNN) and LSTM to simultaneously extract temporal and spatial features. The SE module is an attention mechanism that allows the network to perform feature recalibration, through which the network learns to use global information to selectively emphasize informative features while suppressing less useful features. Moreover, the model uses group convolutions to optimize the number of parameters and reduce the computational complexity. To show the benefit of the proposed model, the performance and efficiency of the improved DeepConvLSTM model and other traditional models are compared on the OPPORTUNITY dataset. In addition, this paper evaluates the impact of hyperparameters on the proposed model and designs an ablation experiment to verify the improvement of each modification. The evaluation results show that the improved DeepConvLSTM model increases the HAR accuracy up to 90.80% given that most computational resources are assigned to more informative features. Moreover, floating-point operations per second (FLOPs) are reduced by 88.93% when the number of group is set to 16 because the group convolution reduces the training parameters greatly.
7 October 2018 in IEEE Communications Letters
In order to blindly detect the modulation order of interference signals in downlink non-orthogonal multiple access systems, a machine learning (ML) algorithm based on Anderson-Darling test is proposed in this letter. The proposed algorithm adopts ML to determine the modulation order of interference user equipment from the raw received constellation points automatically. In feature extraction, a novel feature is introduced to improve the accuracy of blind detection. To evaluate the performance of blind detection, the detection rate and the throughput are simulated under different scenarios. Simulation results show that the proposed algorithm outperforms conventional algorithm on modulation order detection.
6 February 2020 in IEEE Access